freebsd-skq/sbin/newfs/mkfs.c

1147 lines
31 KiB
C
Raw Normal View History

/*
* Copyright (c) 1980, 1989, 1993
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#ifndef lint
#if 0
static char sccsid[] = "@(#)mkfs.c 8.11 (Berkeley) 5/3/95";
#endif
static const char rcsid[] =
1999-08-28 00:22:10 +00:00
"$FreeBSD$";
#endif /* not lint */
#include <err.h>
#include <signal.h>
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#include <unistd.h>
#include <sys/param.h>
#include <sys/time.h>
#include <sys/types.h>
#include <sys/wait.h>
#include <sys/resource.h>
#include <sys/stat.h>
#include <ufs/ufs/dinode.h>
#include <ufs/ufs/dir.h>
#include <ufs/ffs/fs.h>
#include <sys/disklabel.h>
#include <sys/file.h>
#include <sys/mman.h>
#include <sys/ioctl.h>
2002-03-19 20:01:38 +00:00
#include "newfs.h"
/*
* make file system for cylinder-group style file systems
*/
/*
* We limit the size of the inode map to be no more than a
* third of the cylinder group space, since we must leave at
* least an equal amount of space for the block map.
*
* N.B.: MAXIPG must be a multiple of INOPB(fs).
*/
#define MAXIPG(fs) roundup((fs)->fs_bsize * NBBY / 3, INOPB(fs))
#define UMASK 0755
#define MAXINOPB (MAXBSIZE / sizeof(struct dinode))
#define POWEROF2(num) (((num) & ((num) - 1)) == 0)
2002-03-19 20:01:38 +00:00
static union {
struct fs fs;
char pad[SBSIZE];
} fsun;
#define sblock fsun.fs
2002-03-19 20:01:38 +00:00
static struct csum *fscs;
2002-03-19 20:01:38 +00:00
static union {
struct cg cg;
char pad[MAXBSIZE];
} cgun;
#define acg cgun.cg
2002-03-19 20:01:38 +00:00
static struct dinode zino[MAXBSIZE / sizeof(struct dinode)];
2002-03-19 20:01:38 +00:00
static int fsi, fso;
static int randinit;
static daddr_t alloc(int size, int mode);
static long calcipg(long lcpg, long bpcg, off_t *usedbp);
2002-03-19 20:01:38 +00:00
static int charsperline(void);
static void clrblock (struct fs *, unsigned char *, int);
static void fsinit (time_t);
static int ilog2(int);
2002-03-19 20:01:38 +00:00
static void initcg (int, time_t);
static int isblock (struct fs *, unsigned char *, int);
static void iput (struct dinode *, ino_t);
static int makedir (struct direct *, int);
static void rdfs (daddr_t, int, char *);
static void setblock (struct fs *, unsigned char *, int);
static void wtfs (daddr_t, int, char *);
static void wtfsflush (void);
void
mkfs(struct partition *pp, char *fsys, int fi, int fo)
{
long i, mincpc, mincpg, inospercg;
long cylno, j, lwarn = 0;
long used, mincpgcnt, bpcg;
off_t usedb;
long mapcramped, inodecramped;
time_t utime;
quad_t sizepb;
int width;
char tmpbuf[100]; /* XXX this will break in about 2,500 years */
if (Rflag)
utime = 1000000000;
else
time(&utime);
if (!Rflag && !randinit) {
randinit = 1;
1997-06-14 00:17:53 +00:00
srandomdev();
}
fsi = fi;
fso = fo;
sblock.fs_inodefmt = FS_44INODEFMT;
sblock.fs_maxsymlinklen = MAXSYMLINKLEN;
2001-04-03 09:35:36 +00:00
if (Uflag)
sblock.fs_flags |= FS_DOSOFTDEP;
/*
* Validate the given file system size.
* Verify that its last block can actually be accessed.
*/
if (fssize <= 0)
printf("preposterous size %d\n", fssize), exit(13);
wtfs(fssize - (realsectorsize / DEV_BSIZE), realsectorsize,
(char *)&sblock);
/*
* collect and verify the sector and track info
*/
sblock.fs_nsect = secpercyl;
sblock.fs_ntrak = 1;
if (sblock.fs_nsect <= 0)
printf("preposterous nsect %d\n", sblock.fs_nsect), exit(15);
Directory layout preference improvements from Grigoriy Orlov <gluk@ptci.ru>. His description of the problem and solution follow. My own tests show speedups on typical filesystem intensive workloads of 5% to 12% which is very impressive considering the small amount of code change involved. ------ One day I noticed that some file operations run much faster on small file systems then on big ones. I've looked at the ffs algorithms, thought about them, and redesigned the dirpref algorithm. First I want to describe the results of my tests. These results are old and I have improved the algorithm after these tests were done. Nevertheless they show how big the perfomance speedup may be. I have done two file/directory intensive tests on a two OpenBSD systems with old and new dirpref algorithm. The first test is "tar -xzf ports.tar.gz", the second is "rm -rf ports". The ports.tar.gz file is the ports collection from the OpenBSD 2.8 release. It contains 6596 directories and 13868 files. The test systems are: 1. Celeron-450, 128Mb, two IDE drives, the system at wd0, file system for test is at wd1. Size of test file system is 8 Gb, number of cg=991, size of cg is 8m, block size = 8k, fragment size = 1k OpenBSD-current from Dec 2000 with BUFCACHEPERCENT=35 2. PIII-600, 128Mb, two IBM DTLA-307045 IDE drives at i815e, the system at wd0, file system for test is at wd1. Size of test file system is 40 Gb, number of cg=5324, size of cg is 8m, block size = 8k, fragment size = 1k OpenBSD-current from Dec 2000 with BUFCACHEPERCENT=50 You can get more info about the test systems and methods at: http://www.ptci.ru/gluk/dirpref/old/dirpref.html Test Results tar -xzf ports.tar.gz rm -rf ports mode old dirpref new dirpref speedup old dirprefnew dirpref speedup First system normal 667 472 1.41 477 331 1.44 async 285 144 1.98 130 14 9.29 sync 768 616 1.25 477 334 1.43 softdep 413 252 1.64 241 38 6.34 Second system normal 329 81 4.06 263.5 93.5 2.81 async 302 25.7 11.75 112 2.26 49.56 sync 281 57.0 4.93 263 90.5 2.9 softdep 341 40.6 8.4 284 4.76 59.66 "old dirpref" and "new dirpref" columns give a test time in seconds. speedup - speed increasement in times, ie. old dirpref / new dirpref. ------ Algorithm description The old dirpref algorithm is described in comments: /* * Find a cylinder to place a directory. * * The policy implemented by this algorithm is to select from * among those cylinder groups with above the average number of * free inodes, the one with the smallest number of directories. */ A new directory is allocated in a different cylinder groups than its parent directory resulting in a directory tree that is spreaded across all the cylinder groups. This spreading out results in a non-optimal access to the directories and files. When we have a small filesystem it is not a problem but when the filesystem is big then perfomance degradation becomes very apparent. What I mean by a big file system ? 1. A big filesystem is a filesystem which occupy 20-30 or more percent of total drive space, i.e. first and last cylinder are physically located relatively far from each other. 2. It has a relatively large number of cylinder groups, for example more cylinder groups than 50% of the buffers in the buffer cache. The first results in long access times, while the second results in many buffers being used by metadata operations. Such operations use cylinder group blocks and on-disk inode blocks. The cylinder group block (fs->fs_cblkno) contains struct cg, inode and block bit maps. It is 2k in size for the default filesystem parameters. If new and parent directories are located in different cylinder groups then the system performs more input/output operations and uses more buffers. On filesystems with many cylinder groups, lots of cache buffers are used for metadata operations. My solution for this problem is very simple. I allocate many directories in one cylinder group. I also do some things, so that the new allocation method does not cause excessive fragmentation and all directory inodes will not be located at a location far from its file's inodes and data. The algorithm is: /* * Find a cylinder group to place a directory. * * The policy implemented by this algorithm is to allocate a * directory inode in the same cylinder group as its parent * directory, but also to reserve space for its files inodes * and data. Restrict the number of directories which may be * allocated one after another in the same cylinder group * without intervening allocation of files. * * If we allocate a first level directory then force allocation * in another cylinder group. */ My early versions of dirpref give me a good results for a wide range of file operations and different filesystem capacities except one case: those applications that create their entire directory structure first and only later fill this structure with files. My solution for such and similar cases is to limit a number of directories which may be created one after another in the same cylinder group without intervening file creations. For this purpose, I allocate an array of counters at mount time. This array is linked to the superblock fs->fs_contigdirs[cg]. Each time a directory is created the counter increases and each time a file is created the counter decreases. A 60Gb filesystem with 8mb/cg requires 10kb of memory for the counters array. The maxcontigdirs is a maximum number of directories which may be created without an intervening file creation. I found in my tests that the best performance occurs when I restrict the number of directories in one cylinder group such that all its files may be located in the same cylinder group. There may be some deterioration in performance if all the file inodes are in the same cylinder group as its containing directory, but their data partially resides in a different cylinder group. The maxcontigdirs value is calculated to try to prevent this condition. Since there is no way to know how many files and directories will be allocated later I added two optimization parameters in superblock/tunefs. They are: int32_t fs_avgfilesize; /* expected average file size */ int32_t fs_avgfpdir; /* expected # of files per directory */ These parameters have reasonable defaults but may be tweeked for special uses of a filesystem. They are only necessary in rare cases like better tuning a filesystem being used to store a squid cache. I have been using this algorithm for about 3 months. I have done a lot of testing on filesystems with different capacities, average filesize, average number of files per directory, and so on. I think this algorithm has no negative impact on filesystem perfomance. It works better than the default one in all cases. The new dirpref will greatly improve untarring/removing/coping of big directories, decrease load on cvs servers and much more. The new dirpref doesn't speedup a compilation process, but also doesn't slow it down. Obtained from: Grigoriy Orlov <gluk@ptci.ru>
2001-04-10 08:38:59 +00:00
/*
* collect and verify the filesystem density info
*/
sblock.fs_avgfilesize = avgfilesize;
sblock.fs_avgfpdir = avgfilesperdir;
if (sblock.fs_avgfilesize <= 0)
printf("illegal expected average file size %d\n",
sblock.fs_avgfilesize), exit(14);
if (sblock.fs_avgfpdir <= 0)
printf("illegal expected number of files per directory %d\n",
sblock.fs_avgfpdir), exit(15);
/*
* collect and verify the block and fragment sizes
*/
sblock.fs_bsize = bsize;
sblock.fs_fsize = fsize;
if (!POWEROF2(sblock.fs_bsize)) {
printf("block size must be a power of 2, not %d\n",
sblock.fs_bsize);
exit(16);
}
if (!POWEROF2(sblock.fs_fsize)) {
printf("fragment size must be a power of 2, not %d\n",
sblock.fs_fsize);
exit(17);
}
if (sblock.fs_fsize < sectorsize) {
printf("increasing fragment size from %d to sectorsize (%d)\n",
sblock.fs_fsize, sectorsize);
sblock.fs_fsize = sectorsize;
}
if (sblock.fs_bsize < MINBSIZE) {
printf("increasing block size from %d to minimum (%d)\n",
sblock.fs_bsize, MINBSIZE);
sblock.fs_bsize = MINBSIZE;
}
if (sblock.fs_bsize < sblock.fs_fsize) {
printf("increasing block size from %d to fragsize (%d)\n",
sblock.fs_bsize, sblock.fs_fsize);
sblock.fs_bsize = sblock.fs_fsize;
}
if (sblock.fs_fsize * MAXFRAG < sblock.fs_bsize) {
printf("increasing fragsize from %d to block size / %d (%d)\n",
sblock.fs_fsize, MAXFRAG, sblock.fs_bsize / MAXFRAG);
sblock.fs_fsize = sblock.fs_bsize / MAXFRAG;
}
sblock.fs_bmask = ~(sblock.fs_bsize - 1);
sblock.fs_fmask = ~(sblock.fs_fsize - 1);
sblock.fs_qbmask = ~sblock.fs_bmask;
sblock.fs_qfmask = ~sblock.fs_fmask;
sblock.fs_bshift = ilog2(sblock.fs_bsize);
sblock.fs_fshift = ilog2(sblock.fs_fsize);
sblock.fs_frag = numfrags(&sblock, sblock.fs_bsize);
sblock.fs_fragshift = ilog2(sblock.fs_frag);
if (sblock.fs_frag > MAXFRAG) {
printf( "SYSERR: fragsize too small %d (block/frag ratio)\n",
sblock.fs_bsize / MAXFRAG);
exit(21);
}
sblock.fs_nrpos = 1;
sblock.fs_nindir = sblock.fs_bsize / sizeof(daddr_t);
sblock.fs_inopb = sblock.fs_bsize / sizeof(struct dinode);
sblock.fs_nspf = sblock.fs_fsize / sectorsize;
sblock.fs_fsbtodb = ilog2(NSPF(&sblock));
sblock.fs_sblkno =
roundup(howmany(bbsize + sbsize, sblock.fs_fsize), sblock.fs_frag);
sblock.fs_cblkno = (daddr_t)(sblock.fs_sblkno +
roundup(howmany(sbsize, sblock.fs_fsize), sblock.fs_frag));
sblock.fs_iblkno = sblock.fs_cblkno + sblock.fs_frag;
sblock.fs_cgoffset =
roundup(howmany(sblock.fs_nsect, NSPF(&sblock)), sblock.fs_frag);
sblock.fs_cgmask = 0xffffffff;
sblock.fs_maxfilesize = sblock.fs_bsize * NDADDR - 1;
for (sizepb = sblock.fs_bsize, i = 0; i < NIADDR; i++) {
sizepb *= NINDIR(&sblock);
sblock.fs_maxfilesize += sizepb;
}
/*
* Validate specified/determined secpercyl
* and calculate minimum cylinders per group.
*/
sblock.fs_spc = secpercyl;
for (sblock.fs_cpc = NSPB(&sblock), i = sblock.fs_spc;
sblock.fs_cpc > 1 && (i & 1) == 0;
sblock.fs_cpc >>= 1, i >>= 1)
/* void */;
mincpc = sblock.fs_cpc;
bpcg = sblock.fs_spc * sectorsize;
inospercg = roundup(bpcg / sizeof(struct dinode), INOPB(&sblock));
if (inospercg > MAXIPG(&sblock))
inospercg = MAXIPG(&sblock);
used = (sblock.fs_iblkno + inospercg / INOPF(&sblock)) * NSPF(&sblock);
mincpgcnt = howmany(sblock.fs_cgoffset * (~sblock.fs_cgmask) + used,
sblock.fs_spc);
mincpg = roundup(mincpgcnt, mincpc);
/*
* Ensure that cylinder group with mincpg has enough space
* for block maps.
*/
sblock.fs_cpg = mincpg;
sblock.fs_ipg = inospercg;
if (maxcontig > 1)
sblock.fs_contigsumsize = MIN(maxcontig, FS_MAXCONTIG);
mapcramped = 0;
while (CGSIZE(&sblock) > sblock.fs_bsize) {
mapcramped = 1;
if (sblock.fs_bsize < MAXBSIZE) {
sblock.fs_bsize <<= 1;
if ((i & 1) == 0)
i >>= 1;
else {
sblock.fs_cpc <<= 1;
mincpc <<= 1;
mincpg = roundup(mincpgcnt, mincpc);
sblock.fs_cpg = mincpg;
}
sblock.fs_frag <<= 1;
sblock.fs_fragshift += 1;
if (sblock.fs_frag <= MAXFRAG)
continue;
}
if (sblock.fs_fsize == sblock.fs_bsize) {
printf("There is no block size that");
printf(" can support this disk\n");
exit(22);
}
sblock.fs_frag >>= 1;
sblock.fs_fragshift -= 1;
sblock.fs_fsize <<= 1;
sblock.fs_nspf <<= 1;
}
/*
* Ensure that cylinder group with mincpg has enough space for inodes.
*/
inodecramped = 0;
inospercg = calcipg(mincpg, bpcg, &usedb);
sblock.fs_ipg = inospercg;
while (inospercg > MAXIPG(&sblock)) {
inodecramped = 1;
if (mincpc == 1 || sblock.fs_frag == 1 ||
sblock.fs_bsize == MINBSIZE)
break;
printf("With a block size of %d %s %d\n", sblock.fs_bsize,
"minimum bytes per inode is",
(int)((mincpg * (off_t)bpcg - usedb) /
MAXIPG(&sblock) + 1));
sblock.fs_bsize >>= 1;
sblock.fs_frag >>= 1;
sblock.fs_fragshift -= 1;
mincpc >>= 1;
sblock.fs_cpg = roundup(mincpgcnt, mincpc);
if (CGSIZE(&sblock) > sblock.fs_bsize) {
sblock.fs_bsize <<= 1;
break;
}
mincpg = sblock.fs_cpg;
inospercg = calcipg(mincpg, bpcg, &usedb);
sblock.fs_ipg = inospercg;
}
if (inodecramped) {
if (inospercg > MAXIPG(&sblock)) {
printf("Minimum bytes per inode is %d\n",
(int)((mincpg * (off_t)bpcg - usedb) /
MAXIPG(&sblock) + 1));
} else if (!mapcramped) {
printf("With %d bytes per inode, ", density);
1998-06-28 20:11:23 +00:00
printf("minimum cylinders per group is %ld\n", mincpg);
}
}
if (mapcramped) {
printf("With %d sectors per cylinder, ", sblock.fs_spc);
1998-06-28 20:11:23 +00:00
printf("minimum cylinders per group is %ld\n", mincpg);
}
if (inodecramped || mapcramped) {
if (sblock.fs_bsize != bsize)
printf("%s to be changed from %d to %d\n",
"This requires the block size",
bsize, sblock.fs_bsize);
if (sblock.fs_fsize != fsize)
printf("\t%s to be changed from %d to %d\n",
"and the fragment size", fsize, sblock.fs_fsize);
exit(23);
}
1995-05-30 06:12:45 +00:00
/*
* Calculate the number of cylinders per group
*/
sblock.fs_cpg = cpg;
if (sblock.fs_cpg % mincpc != 0) {
1998-06-28 20:11:23 +00:00
printf("%s groups must have a multiple of %ld cylinders\n",
cpgflg ? "Cylinder" : "Warning: cylinder", mincpc);
sblock.fs_cpg = roundup(sblock.fs_cpg, mincpc);
if (!cpgflg)
cpg = sblock.fs_cpg;
}
/*
* Must ensure there is enough space for inodes.
*/
sblock.fs_ipg = calcipg(sblock.fs_cpg, bpcg, &usedb);
while (sblock.fs_ipg > MAXIPG(&sblock)) {
inodecramped = 1;
sblock.fs_cpg -= mincpc;
sblock.fs_ipg = calcipg(sblock.fs_cpg, bpcg, &usedb);
}
/*
* Must ensure there is enough space to hold block map.
*/
while (CGSIZE(&sblock) > sblock.fs_bsize) {
mapcramped = 1;
sblock.fs_cpg -= mincpc;
sblock.fs_ipg = calcipg(sblock.fs_cpg, bpcg, &usedb);
}
sblock.fs_fpg = (sblock.fs_cpg * sblock.fs_spc) / NSPF(&sblock);
if ((sblock.fs_cpg * sblock.fs_spc) % NSPB(&sblock) != 0) {
1998-06-28 20:11:23 +00:00
printf("panic (fs_cpg * fs_spc) %% NSPF != 0");
exit(24);
}
if (sblock.fs_cpg < mincpg) {
1998-06-28 20:11:23 +00:00
printf("cylinder groups must have at least %ld cylinders\n",
mincpg);
exit(25);
} else if (sblock.fs_cpg != cpg) {
if (!cpgflg)
printf("Warning: ");
else if (!mapcramped && !inodecramped)
exit(26);
if (mapcramped && inodecramped)
printf("Block size and bytes per inode restrict");
else if (mapcramped)
printf("Block size restricts");
else
printf("Bytes per inode restrict");
printf(" cylinders per group to %d.\n", sblock.fs_cpg);
if (cpgflg)
exit(27);
}
sblock.fs_cgsize = fragroundup(&sblock, CGSIZE(&sblock));
/*
* Now have size for file system and nsect and ntrak.
* Determine number of cylinders and blocks in the file system.
*/
sblock.fs_size = fssize = dbtofsb(&sblock, fssize);
sblock.fs_ncyl = fssize * NSPF(&sblock) / sblock.fs_spc;
if (fssize * NSPF(&sblock) > sblock.fs_ncyl * sblock.fs_spc) {
sblock.fs_ncyl++;
lwarn = 1;
}
if (sblock.fs_ncyl < 1) {
printf("file systems must have at least one cylinder\n");
exit(28);
}
/*
* Determine feasability/values of rotational layout tables.
*
* The size of the rotational layout tables is limited by the
* size of the superblock, SBSIZE. The amount of space available
* for tables is calculated as (SBSIZE - sizeof (struct fs)).
* The size of these tables is inversely proportional to the block
* size of the file system. The size increases if sectors per track
* are not powers of two, because more cylinders must be described
* by the tables before the rotational pattern repeats (fs_cpc).
*/
sblock.fs_interleave = 1;
sblock.fs_trackskew = 0;
sblock.fs_npsect = secpercyl;
sblock.fs_postblformat = FS_DYNAMICPOSTBLFMT;
sblock.fs_sbsize = fragroundup(&sblock, sizeof(struct fs));
if (sblock.fs_sbsize > SBSIZE)
sblock.fs_sbsize = SBSIZE;
sblock.fs_cpc = 0;
/*
* Compute/validate number of cylinder groups.
*/
sblock.fs_ncg = sblock.fs_ncyl / sblock.fs_cpg;
if (sblock.fs_ncyl % sblock.fs_cpg)
sblock.fs_ncg++;
sblock.fs_dblkno = sblock.fs_iblkno + sblock.fs_ipg / INOPF(&sblock);
i = MIN(~sblock.fs_cgmask, sblock.fs_ncg - 1);
if (cgdmin(&sblock, i) - cgbase(&sblock, i) >= sblock.fs_fpg) {
1998-06-28 20:11:23 +00:00
printf("inode blocks/cyl group (%ld) >= data blocks (%ld)\n",
cgdmin(&sblock, i) - cgbase(&sblock, i) / sblock.fs_frag,
1998-06-28 20:11:23 +00:00
(long)(sblock.fs_fpg / sblock.fs_frag));
printf("number of cylinders per cylinder group (%d) %s.\n",
sblock.fs_cpg, "must be increased");
exit(29);
}
j = sblock.fs_ncg - 1;
if ((i = fssize - j * sblock.fs_fpg) < sblock.fs_fpg &&
cgdmin(&sblock, j) - cgbase(&sblock, j) > i) {
if (j == 0) {
printf("Filesystem must have at least %d sectors\n",
NSPF(&sblock) *
(cgdmin(&sblock, 0) + 3 * sblock.fs_frag));
exit(30);
}
printf(
"Warning: inode blocks/cyl group (%ld) >= data blocks (%ld) in last\n",
(cgdmin(&sblock, j) - cgbase(&sblock, j)) / sblock.fs_frag,
i / sblock.fs_frag);
printf(
" cylinder group. This implies %ld sector(s) cannot be allocated.\n",
i * NSPF(&sblock));
sblock.fs_ncg--;
sblock.fs_ncyl -= sblock.fs_ncyl % sblock.fs_cpg;
sblock.fs_size = fssize = sblock.fs_ncyl * sblock.fs_spc /
NSPF(&sblock);
lwarn = 0;
}
if (lwarn) {
printf("Warning: %d sector(s) in last cylinder unallocated\n",
sblock.fs_spc -
(fssize * NSPF(&sblock) - (sblock.fs_ncyl - 1) *
sblock.fs_spc));
}
/*
* fill in remaining fields of the super block
*/
sblock.fs_csaddr = cgdmin(&sblock, 0);
sblock.fs_cssize =
fragroundup(&sblock, sblock.fs_ncg * sizeof(struct csum));
/*
* The superblock fields 'fs_csmask' and 'fs_csshift' are no
* longer used. However, we still initialise them so that the
* filesystem remains compatible with old kernels.
*/
i = sblock.fs_bsize / sizeof(struct csum);
sblock.fs_csmask = ~(i - 1);
sblock.fs_csshift = ilog2(i);
fscs = (struct csum *)calloc(1, sblock.fs_cssize);
if (fscs == NULL)
errx(31, "calloc failed");
sblock.fs_magic = FS_MAGIC;
sblock.fs_rotdelay = 0;
sblock.fs_minfree = minfree;
sblock.fs_maxcontig = maxcontig;
sblock.fs_maxbpg = maxbpg;
sblock.fs_rps = 60;
sblock.fs_optim = opt;
sblock.fs_cgrotor = 0;
sblock.fs_cstotal.cs_ndir = 0;
sblock.fs_cstotal.cs_nbfree = 0;
sblock.fs_cstotal.cs_nifree = 0;
sblock.fs_cstotal.cs_nffree = 0;
sblock.fs_fmod = 0;
sblock.fs_ronly = 0;
1994-08-26 10:41:13 +00:00
sblock.fs_clean = 1;
sblock.fs_id[0] = (long)utime;
sblock.fs_id[1] = random();
/*
* Dump out summary information about file system.
*/
printf("%s:\t%d sectors in %d %s of %d tracks, %d sectors\n",
fsys, sblock.fs_size * NSPF(&sblock), sblock.fs_ncyl,
"cylinders", sblock.fs_ntrak, sblock.fs_nsect);
#define B2MBFACTOR (1 / (1024.0 * 1024.0))
printf("\t%.1fMB in %d cyl groups (%d c/g, %.2fMB/g, %d i/g)%s\n",
(float)sblock.fs_size * sblock.fs_fsize * B2MBFACTOR,
sblock.fs_ncg, sblock.fs_cpg,
(float)sblock.fs_fpg * sblock.fs_fsize * B2MBFACTOR,
sblock.fs_ipg,
sblock.fs_flags & FS_DOSOFTDEP ? " SOFTUPDATES" : "");
#undef B2MBFACTOR
/*
* Now build the cylinders group blocks and
* then print out indices of cylinder groups.
*/
printf("super-block backups (for fsck -b #) at:\n");
i = 0;
width = charsperline();
for (cylno = 0; cylno < sblock.fs_ncg; cylno++) {
initcg(cylno, utime);
j = snprintf(tmpbuf, sizeof(tmpbuf), " %ld%s",
fsbtodb(&sblock, cgsblock(&sblock, cylno)),
cylno < (sblock.fs_ncg-1) ? "," : "");
if (j < 0)
tmpbuf[j = 0] = '\0';
1998-06-28 20:11:23 +00:00
if (i + j >= width) {
printf("\n");
i = 0;
}
i += j;
printf("%s", tmpbuf);
fflush(stdout);
}
printf("\n");
if (Nflag)
exit(0);
/*
* Now construct the initial file system,
* then write out the super-block.
*/
fsinit(utime);
sblock.fs_time = utime;
wtfs((int)SBOFF / sectorsize, sbsize, (char *)&sblock);
for (i = 0; i < sblock.fs_cssize; i += sblock.fs_bsize)
wtfs(fsbtodb(&sblock, sblock.fs_csaddr + numfrags(&sblock, i)),
sblock.fs_cssize - i < sblock.fs_bsize ?
sblock.fs_cssize - i : sblock.fs_bsize,
((char *)fscs) + i);
1995-05-30 06:12:45 +00:00
/*
* Write out the duplicate super blocks
*/
for (cylno = 0; cylno < sblock.fs_ncg; cylno++)
wtfs(fsbtodb(&sblock, cgsblock(&sblock, cylno)),
sbsize, (char *)&sblock);
wtfsflush();
/*
* Update information about this partion in pack
* label, to that it may be updated on disk.
*/
pp->p_fstype = FS_BSDFFS;
pp->p_fsize = sblock.fs_fsize;
pp->p_frag = sblock.fs_frag;
pp->p_cpg = sblock.fs_cpg;
}
/*
* Initialize a cylinder group.
*/
void
initcg(int cylno, time_t utime)
{
daddr_t cbase, d, dlower, dupper, dmax, blkno;
struct csum *cs;
long i, j;
/*
* Determine block bounds for cylinder group.
* Allow space for super block summary information in first
* cylinder group.
*/
cbase = cgbase(&sblock, cylno);
dmax = cbase + sblock.fs_fpg;
if (dmax > sblock.fs_size)
dmax = sblock.fs_size;
dlower = cgsblock(&sblock, cylno) - cbase;
dupper = cgdmin(&sblock, cylno) - cbase;
if (cylno == 0)
dupper += howmany(sblock.fs_cssize, sblock.fs_fsize);
cs = fscs + cylno;
memset(&acg, 0, sblock.fs_cgsize);
acg.cg_time = utime;
acg.cg_magic = CG_MAGIC;
acg.cg_cgx = cylno;
if (cylno == sblock.fs_ncg - 1)
acg.cg_ncyl = sblock.fs_ncyl % sblock.fs_cpg;
else
acg.cg_ncyl = sblock.fs_cpg;
acg.cg_niblk = sblock.fs_ipg;
acg.cg_ndblk = dmax - cbase;
if (sblock.fs_contigsumsize > 0)
acg.cg_nclusterblks = acg.cg_ndblk / sblock.fs_frag;
acg.cg_btotoff = &acg.cg_space[0] - (u_char *)(&acg.cg_firstfield);
acg.cg_boff = acg.cg_btotoff + sblock.fs_cpg * sizeof(int32_t);
1995-05-30 06:12:45 +00:00
acg.cg_iusedoff = acg.cg_boff +
sblock.fs_cpg * sizeof(u_int16_t);
acg.cg_freeoff = acg.cg_iusedoff + howmany(sblock.fs_ipg, NBBY);
if (sblock.fs_contigsumsize <= 0) {
acg.cg_nextfreeoff = acg.cg_freeoff +
howmany(sblock.fs_cpg * sblock.fs_spc / NSPF(&sblock),
NBBY);
} else {
acg.cg_clustersumoff = acg.cg_freeoff + howmany
(sblock.fs_cpg * sblock.fs_spc / NSPF(&sblock), NBBY) -
sizeof(u_int32_t);
acg.cg_clustersumoff =
roundup(acg.cg_clustersumoff, sizeof(u_int32_t));
acg.cg_clusteroff = acg.cg_clustersumoff +
(sblock.fs_contigsumsize + 1) * sizeof(u_int32_t);
acg.cg_nextfreeoff = acg.cg_clusteroff + howmany
(sblock.fs_cpg * sblock.fs_spc / NSPB(&sblock), NBBY);
}
if (acg.cg_nextfreeoff - (long)(&acg.cg_firstfield) >
sblock.fs_cgsize) {
printf("Panic: cylinder group too big\n");
exit(37);
}
acg.cg_cs.cs_nifree += sblock.fs_ipg;
if (cylno == 0)
for (i = 0; i < ROOTINO; i++) {
setbit(cg_inosused(&acg), i);
acg.cg_cs.cs_nifree--;
}
for (i = 0; i < sblock.fs_ipg / INOPF(&sblock); i += sblock.fs_frag) {
for (j = 0; j < sblock.fs_bsize / sizeof(struct dinode); j++)
zino[j].di_gen = random();
wtfs(fsbtodb(&sblock, cgimin(&sblock, cylno) + i),
sblock.fs_bsize, (char *)zino);
}
if (cylno > 0) {
/*
* In cylno 0, beginning space is reserved
* for boot and super blocks.
*/
for (d = 0; d < dlower; d += sblock.fs_frag) {
blkno = d / sblock.fs_frag;
setblock(&sblock, cg_blksfree(&acg), blkno);
if (sblock.fs_contigsumsize > 0)
setbit(cg_clustersfree(&acg), blkno);
acg.cg_cs.cs_nbfree++;
cg_blktot(&acg)[cbtocylno(&sblock, d)]++;
cg_blks(&sblock, &acg, cbtocylno(&sblock, d))
[cbtorpos(&sblock, d)]++;
}
sblock.fs_dsize += dlower;
}
sblock.fs_dsize += acg.cg_ndblk - dupper;
if ((i = dupper % sblock.fs_frag)) {
acg.cg_frsum[sblock.fs_frag - i]++;
for (d = dupper + sblock.fs_frag - i; dupper < d; dupper++) {
setbit(cg_blksfree(&acg), dupper);
acg.cg_cs.cs_nffree++;
}
}
for (d = dupper; d + sblock.fs_frag <= dmax - cbase;) {
blkno = d / sblock.fs_frag;
setblock(&sblock, cg_blksfree(&acg), blkno);
if (sblock.fs_contigsumsize > 0)
setbit(cg_clustersfree(&acg), blkno);
acg.cg_cs.cs_nbfree++;
cg_blktot(&acg)[cbtocylno(&sblock, d)]++;
cg_blks(&sblock, &acg, cbtocylno(&sblock, d))
[cbtorpos(&sblock, d)]++;
d += sblock.fs_frag;
}
if (d < dmax - cbase) {
acg.cg_frsum[dmax - cbase - d]++;
for (; d < dmax - cbase; d++) {
setbit(cg_blksfree(&acg), d);
acg.cg_cs.cs_nffree++;
}
}
if (sblock.fs_contigsumsize > 0) {
int32_t *sump = cg_clustersum(&acg);
u_char *mapp = cg_clustersfree(&acg);
int map = *mapp++;
int bit = 1;
int run = 0;
for (i = 0; i < acg.cg_nclusterblks; i++) {
if ((map & bit) != 0)
run++;
else if (run != 0) {
if (run > sblock.fs_contigsumsize)
run = sblock.fs_contigsumsize;
sump[run]++;
run = 0;
}
if ((i & (NBBY - 1)) != NBBY - 1)
bit <<= 1;
else {
map = *mapp++;
bit = 1;
}
}
if (run != 0) {
if (run > sblock.fs_contigsumsize)
run = sblock.fs_contigsumsize;
sump[run]++;
}
}
sblock.fs_cstotal.cs_ndir += acg.cg_cs.cs_ndir;
sblock.fs_cstotal.cs_nffree += acg.cg_cs.cs_nffree;
sblock.fs_cstotal.cs_nbfree += acg.cg_cs.cs_nbfree;
sblock.fs_cstotal.cs_nifree += acg.cg_cs.cs_nifree;
*cs = acg.cg_cs;
wtfs(fsbtodb(&sblock, cgtod(&sblock, cylno)),
sblock.fs_bsize, (char *)&acg);
}
/*
* initialize the file system
*/
struct dinode node;
#ifdef LOSTDIR
#define PREDEFDIR 3
#else
#define PREDEFDIR 2
#endif
struct direct root_dir[] = {
{ ROOTINO, sizeof(struct direct), DT_DIR, 1, "." },
{ ROOTINO, sizeof(struct direct), DT_DIR, 2, ".." },
#ifdef LOSTDIR
{ LOSTFOUNDINO, sizeof(struct direct), DT_DIR, 10, "lost+found" },
#endif
};
struct odirect {
u_long d_ino;
u_short d_reclen;
u_short d_namlen;
u_char d_name[MAXNAMLEN + 1];
} oroot_dir[] = {
{ ROOTINO, sizeof(struct direct), 1, "." },
{ ROOTINO, sizeof(struct direct), 2, ".." },
#ifdef LOSTDIR
{ LOSTFOUNDINO, sizeof(struct direct), 10, "lost+found" },
#endif
};
#ifdef LOSTDIR
struct direct lost_found_dir[] = {
{ LOSTFOUNDINO, sizeof(struct direct), DT_DIR, 1, "." },
{ ROOTINO, sizeof(struct direct), DT_DIR, 2, ".." },
{ 0, DIRBLKSIZ, 0, 0, 0 },
};
struct odirect olost_found_dir[] = {
{ LOSTFOUNDINO, sizeof(struct direct), 1, "." },
{ ROOTINO, sizeof(struct direct), 2, ".." },
{ 0, DIRBLKSIZ, 0, 0 },
};
#endif
char buf[MAXBSIZE];
void
fsinit(time_t utime)
{
#ifdef LOSTDIR
int i;
#endif
/*
* initialize the node
*/
node.di_atime = utime;
node.di_mtime = utime;
node.di_ctime = utime;
#ifdef LOSTDIR
/*
* create the lost+found directory
*/
(void)makedir(lost_found_dir, 2);
for (i = DIRBLKSIZ; i < sblock.fs_bsize; i += DIRBLKSIZ)
memmove(&buf[i], &lost_found_dir[2],
DIRSIZ(0, &lost_found_dir[2]));
node.di_mode = IFDIR | UMASK;
node.di_nlink = 2;
node.di_size = sblock.fs_bsize;
node.di_db[0] = alloc(node.di_size, node.di_mode);
node.di_blocks = btodb(fragroundup(&sblock, node.di_size));
wtfs(fsbtodb(&sblock, node.di_db[0]), node.di_size, buf);
iput(&node, LOSTFOUNDINO);
#endif
/*
* create the root directory
*/
node.di_mode = IFDIR | UMASK;
node.di_nlink = PREDEFDIR;
node.di_size = makedir(root_dir, PREDEFDIR);
node.di_db[0] = alloc(sblock.fs_fsize, node.di_mode);
node.di_blocks = btodb(fragroundup(&sblock, node.di_size));
wtfs(fsbtodb(&sblock, node.di_db[0]), sblock.fs_fsize, buf);
iput(&node, ROOTINO);
}
/*
* construct a set of directory entries in "buf".
* return size of directory.
*/
int
makedir(struct direct *protodir, int entries)
{
char *cp;
int i, spcleft;
spcleft = DIRBLKSIZ;
for (cp = buf, i = 0; i < entries - 1; i++) {
protodir[i].d_reclen = DIRSIZ(0, &protodir[i]);
memmove(cp, &protodir[i], protodir[i].d_reclen);
cp += protodir[i].d_reclen;
spcleft -= protodir[i].d_reclen;
}
protodir[i].d_reclen = spcleft;
memmove(cp, &protodir[i], DIRSIZ(0, &protodir[i]));
return (DIRBLKSIZ);
}
/*
* allocate a block or frag
*/
daddr_t
alloc(int size, int mode)
{
int i, frag;
daddr_t d, blkno;
rdfs(fsbtodb(&sblock, cgtod(&sblock, 0)), sblock.fs_cgsize,
(char *)&acg);
if (acg.cg_magic != CG_MAGIC) {
printf("cg 0: bad magic number\n");
return (0);
}
if (acg.cg_cs.cs_nbfree == 0) {
printf("first cylinder group ran out of space\n");
return (0);
}
for (d = 0; d < acg.cg_ndblk; d += sblock.fs_frag)
if (isblock(&sblock, cg_blksfree(&acg), d / sblock.fs_frag))
goto goth;
printf("internal error: can't find block in cyl 0\n");
return (0);
goth:
blkno = fragstoblks(&sblock, d);
clrblock(&sblock, cg_blksfree(&acg), blkno);
if (sblock.fs_contigsumsize > 0)
clrbit(cg_clustersfree(&acg), blkno);
acg.cg_cs.cs_nbfree--;
sblock.fs_cstotal.cs_nbfree--;
fscs[0].cs_nbfree--;
if (mode & IFDIR) {
acg.cg_cs.cs_ndir++;
sblock.fs_cstotal.cs_ndir++;
fscs[0].cs_ndir++;
}
cg_blktot(&acg)[cbtocylno(&sblock, d)]--;
cg_blks(&sblock, &acg, cbtocylno(&sblock, d))[cbtorpos(&sblock, d)]--;
if (size != sblock.fs_bsize) {
frag = howmany(size, sblock.fs_fsize);
fscs[0].cs_nffree += sblock.fs_frag - frag;
sblock.fs_cstotal.cs_nffree += sblock.fs_frag - frag;
acg.cg_cs.cs_nffree += sblock.fs_frag - frag;
acg.cg_frsum[sblock.fs_frag - frag]++;
for (i = frag; i < sblock.fs_frag; i++)
setbit(cg_blksfree(&acg), d + i);
}
wtfs(fsbtodb(&sblock, cgtod(&sblock, 0)), sblock.fs_cgsize,
(char *)&acg);
return (d);
}
/*
* Calculate number of inodes per group.
*/
long
calcipg(long lcpg, long bpcg, off_t *usedbp)
{
int i;
long ipg, new_ipg, ncg, ncyl;
off_t usedb;
/*
* Prepare to scale by fssize / (number of sectors in cylinder groups).
* Note that fssize is still in sectors, not filesystem blocks.
*/
ncyl = howmany(fssize, (u_int)secpercyl);
ncg = howmany(ncyl, lcpg);
/*
* Iterate a few times to allow for ipg depending on itself.
*/
ipg = 0;
for (i = 0; i < 10; i++) {
usedb = (sblock.fs_iblkno + ipg / INOPF(&sblock)) *
NSPF(&sblock) * (off_t)sectorsize;
new_ipg = (lcpg * (quad_t)bpcg - usedb) / density *
fssize / ncg / secpercyl / lcpg;
new_ipg = roundup(new_ipg, INOPB(&sblock));
if (new_ipg == ipg)
break;
ipg = new_ipg;
}
*usedbp = usedb;
return (ipg);
}
/*
* Allocate an inode on the disk
*/
void
iput(struct dinode *ip, ino_t ino)
{
struct dinode lbuf[MAXINOPB];
daddr_t d;
int c;
ip->di_gen = random();
c = ino_to_cg(&sblock, ino);
rdfs(fsbtodb(&sblock, cgtod(&sblock, 0)), sblock.fs_cgsize,
(char *)&acg);
if (acg.cg_magic != CG_MAGIC) {
printf("cg 0: bad magic number\n");
exit(31);
}
acg.cg_cs.cs_nifree--;
setbit(cg_inosused(&acg), ino);
wtfs(fsbtodb(&sblock, cgtod(&sblock, 0)), sblock.fs_cgsize,
(char *)&acg);
sblock.fs_cstotal.cs_nifree--;
fscs[0].cs_nifree--;
if (ino >= sblock.fs_ipg * sblock.fs_ncg) {
printf("fsinit: inode value out of range (%d).\n", ino);
exit(32);
}
d = fsbtodb(&sblock, ino_to_fsba(&sblock, ino));
rdfs(d, sblock.fs_bsize, (char *)lbuf);
lbuf[ino_to_fsbo(&sblock, ino)] = *ip;
wtfs(d, sblock.fs_bsize, (char *)lbuf);
}
/*
* read a block from the file system
*/
void
rdfs(daddr_t bno, int size, char *bf)
{
int n;
wtfsflush();
if (lseek(fsi, (off_t)bno * sectorsize, 0) < 0) {
1998-06-28 20:11:23 +00:00
printf("seek error: %ld\n", (long)bno);
err(33, "rdfs");
}
n = read(fsi, bf, size);
if (n != size) {
1998-06-28 20:11:23 +00:00
printf("read error: %ld\n", (long)bno);
err(34, "rdfs");
}
}
#define WCSIZE (128 * 1024)
daddr_t wc_sect; /* units of sectorsize */
int wc_end; /* bytes */
static char wc[WCSIZE]; /* bytes */
/*
* Flush dirty write behind buffer.
*/
static void
wtfsflush()
{
int n;
if (wc_end) {
if (lseek(fso, (off_t)wc_sect * sectorsize, SEEK_SET) < 0) {
printf("seek error: %ld\n", (long)wc_sect);
err(35, "wtfs - writecombine");
}
n = write(fso, wc, wc_end);
if (n != wc_end) {
printf("write error: %ld\n", (long)wc_sect);
err(36, "wtfs - writecombine");
}
wc_end = 0;
}
}
/*
* write a block to the file system
*/
static void
wtfs(daddr_t bno, int size, char *bf)
{
int done, n;
if (Nflag)
return;
done = 0;
if (wc_end == 0 && size <= WCSIZE) {
wc_sect = bno;
bcopy(bf, wc, size);
wc_end = size;
if (wc_end < WCSIZE)
return;
done = 1;
}
if ((off_t)wc_sect * sectorsize + wc_end == (off_t)bno * sectorsize &&
wc_end + size <= WCSIZE) {
bcopy(bf, wc + wc_end, size);
wc_end += size;
if (wc_end < WCSIZE)
return;
done = 1;
}
wtfsflush();
if (done)
return;
if (lseek(fso, (off_t)bno * sectorsize, SEEK_SET) < 0) {
1998-06-28 20:11:23 +00:00
printf("seek error: %ld\n", (long)bno);
err(35, "wtfs");
}
n = write(fso, bf, size);
if (n != size) {
1998-06-28 20:11:23 +00:00
printf("write error: %ld\n", (long)bno);
err(36, "wtfs");
}
}
/*
* check if a block is available
*/
static int
isblock(struct fs *fs, unsigned char *cp, int h)
{
unsigned char mask;
switch (fs->fs_frag) {
case 8:
return (cp[h] == 0xff);
case 4:
mask = 0x0f << ((h & 0x1) << 2);
return ((cp[h >> 1] & mask) == mask);
case 2:
mask = 0x03 << ((h & 0x3) << 1);
return ((cp[h >> 2] & mask) == mask);
case 1:
mask = 0x01 << (h & 0x7);
return ((cp[h >> 3] & mask) == mask);
default:
fprintf(stderr, "isblock bad fs_frag %d\n", fs->fs_frag);
return (0);
}
}
/*
* take a block out of the map
*/
static void
clrblock(struct fs *fs, unsigned char *cp, int h)
{
switch ((fs)->fs_frag) {
case 8:
cp[h] = 0;
return;
case 4:
cp[h >> 1] &= ~(0x0f << ((h & 0x1) << 2));
return;
case 2:
cp[h >> 2] &= ~(0x03 << ((h & 0x3) << 1));
return;
case 1:
cp[h >> 3] &= ~(0x01 << (h & 0x7));
return;
default:
fprintf(stderr, "clrblock bad fs_frag %d\n", fs->fs_frag);
return;
}
}
/*
* put a block into the map
*/
static void
setblock(struct fs *fs, unsigned char *cp, int h)
{
switch (fs->fs_frag) {
case 8:
cp[h] = 0xff;
return;
case 4:
cp[h >> 1] |= (0x0f << ((h & 0x1) << 2));
return;
case 2:
cp[h >> 2] |= (0x03 << ((h & 0x3) << 1));
return;
case 1:
cp[h >> 3] |= (0x01 << (h & 0x7));
return;
default:
fprintf(stderr, "setblock bad fs_frag %d\n", fs->fs_frag);
return;
}
}
/*
* Determine the number of characters in a
* single line.
*/
static int
2002-03-19 20:01:38 +00:00
charsperline(void)
{
int columns;
char *cp;
struct winsize ws;
columns = 0;
if (ioctl(0, TIOCGWINSZ, &ws) != -1)
columns = ws.ws_col;
if (columns == 0 && (cp = getenv("COLUMNS")))
columns = atoi(cp);
if (columns == 0)
columns = 80; /* last resort */
return (columns);
}
static int
ilog2(int val)
{
u_int n;
for (n = 0; n < sizeof(n) * NBBY; n++)
if (1 << n == val)
return (n);
errx(1, "ilog2: %d is not a power of 2\n", val);
}